Sensitive H2 sensors at low concentrations and room temperature are desired for the early warning and control of hydrogen leakage. In this paper, a resistive sensor based on Pt-doped In2O3 nanoparticles was fabricated using inkjet printing process. The H2 sensing performance of the sensor was evaluated at low concentrations below 1% at room temperature. It exhibited a relative high response of 42.34% to 0.6% H2. As the relative humidity of 0.5% H2 decreased from 34% to 23%, the response decreased slightly from 34% to 23%. The sensing principle and the humidity effect were discussed. A dynamic current sensing model for dry H2 detection was proposed based on Wolkenstein theory and experimentally verified to be able to predict the sensing behavior of the sensor. The H2 concentration can be calculated within a short measurement time using the model without waiting for the saturation of the response, which significantly reduces the sensing and recovery time of the sensor. The sensor is expected to be a promising candidate for room-temperature H2 detection, and the proposed model could be very helpful in promoting the application of the sensor for real-time H2 leakage monitoring.
Humidity and oxygen have significant impacts on the accuracy of hydrogen detection, especially for metal oxide semiconductor sensors at room temperature. Addressing this challenge, this study employs a screen-printed 1 × 2 resistive sensor array made from an identical 1 wt.% platinum-modified tin oxide nanoparticle material. Fabrication variability between the two sensing elements was intentionally leveraged to enhance array output differentiation and information content. Systematic hydrogen-sensing tests were conducted on the array under diverse oxygen and moisture conditions. Three distinct feature types—the steady-state value, resistance change, and area under the curve—were extracted from the output of each array element. These features, integrated with their quotient, formed a nine-feature vector matrix. A multiple linear regression model based on this array output was developed and validated for hydrogen prediction, achieving a coefficient of determination of 0.95, a mean absolute error of 125 ppm, and a mean relative standard deviation of 7.07%. The combined information of the array provided significantly more stable and precise hydrogen concentration predictions than linear or nonlinear models based on individual sensor features. This approach offers a promising path for mass-producing highly interference-resistant, precise, and stable room-temperature hydrogen sensor arrays.
Iron concentrate grade acts as a critical control indicator in iron ore mining and beneficiation, and on-line analysis facilitates stable product quality and improved metal recovery. Conventional detection methods suffer from severe time lag and fail to meet real-time monitoring requirements. Laser-induced breakdown spectroscopy (LIBS) enables on-line grade measurement of iron ore slurry; nevertheless, single-spectrum modeling is susceptible to matrix effects induced by particle size variation. Existing spectrum-particle size fusion models lack sufficient cross-feature interaction, resulting in limited prediction accuracy. To address these limitations, this paper proposes a partial least squares (PLS)-guided bidirectional attention fusion network (PLS-BAF-Net) for LIBS spectra and particle size signals. The proposed model takes interpretable spectral features as prior knowledge and leverages the feature interaction capability of the bidirectional attention mechanism to effectively mitigate particle size-induced matrix interference. Field data acquisition and practical validation are implemented using a self-developed LIBS slurry detection system denoted SIA-LIBSlurry, followed by comprehensive comparative experiments. The experimental results reveal that compared with the conventional PLS model, the proposed network achieves a 0.626 (41.4%) reduction in RMSE, a 0.375 (35.01%) reduction in MAE, and an R2 improvement of 0.188 (26.33%). Its overall predictive performance outperforms data-level, feature-level and decision-level fusion models based on PLS. The proposed method remarkably enhances the on-line quantitative detection accuracy of iron grade in iron ore slurry, and satisfies the real-time on-line monitoring requirements for industrial iron ore slurry production.